The Long Cycle in International Relations: A Markov Specification

Despite mounting empirical evidence against cycles, there is still a considerable literature in international relations supporting the claim of a Kondratieff wave in war. But the empirical evidence against cycles, however, rests on tests that are inappropriate for a nonlinear process where the cycle recurs at irregular intervals. So, the claim of a long cycle has yet to be falsified. This paper attempts to make an appropriate empirical test of the long cycle hypothesis. I specify and test the war cycle as segmented time trends under Markov assumptions. The results show that there is a cycle in war. The swings are not Kondratieffs, nor are they generated by the accumulation of random error. Instead, I find evidence for a war-price cycle of approximately 21 years duration. I conclude that most likely the war cycle reflects Kuznets rather than Kondratieff processes at work.